Fatality census · NHTSA FARS

Every fatal crash in the United States, 1975–2024

A complete census of 1,872,464 fatal motor-vehicle crashes and 2,080,576 deaths recorded by NHTSA's Fatality Analysis Reporting System over 50 years. One record per crash, with every coded field resolved to a label. Federal Highway Administration travel data is included, so any measure can be expressed as a rate.

0.01.83.5198020012022
Figure 1. United States road-death rate, 1980–2022: deaths per 100 million vehicle-miles travelled. The rate fell about 2.5-fold as total travel rose. Deaths, NHTSA FARS; travel, FHWA.
deaths
2,080,576
fatal crashes
1,872,464
years on record
50
deaths / 100M mi · 2022
1.34

Each record carries the crash location, date and time, road class and weather, every vehicle and person involved, and coded contributing factors including speeding, alcohol, drug involvement, distraction, and vehicle type.

Figure 2. Occupant deaths per 1,000 involved vehicles in two-vehicle fatal crashes, by vehicle type (2015–2024). In car-to-light-truck collisions, car occupants die 3.0 times as often as light-truck occupants (29,322 vs 9,659 deaths); against heavy trucks, 31 times. Vehicle type is classified from the FARS body-type code for 1991 onward (99% of involved vehicles).

Use the data

Example analyses

Each link opens the Explorer or record browser with a configuration applied. Charts are computed in the browser against the full census, and the view URL is shareable.

For machines · crash_stats.json · FARS query DB (gzip) · sources NHTSA FARS + FHWA exposure (VMT, drivers, population)

The NHTSA Fatality Analysis Reporting System (FARS) is a census of every fatal crash on a US public road, deduplicated and decoded by an autonomous pipeline. The most recent FARS year is an Initial Release (preliminary), refreshed when finalized. Rate baselines normalize the counts by FHWA travel-exposure data: vehicle-miles travelled (1980–2022) and licensed drivers and resident population (1998–2021).